This paper presents a new way to design mobility assistant devices that help people stand up from a seated position, especially those suffering from neuro-muscular disabilities and challenges. The researchers developed a method using machine learning and kinematics (the study of motion) to create six-bar linkage mechanisms that mimic the natural movement of the human body during standing. Unlike traditional design approaches, which can be slow and limited, their method uses deep learning to quickly generate and evaluate thousands of possible conceptual designs. By training a neural network to recognize useful movement patterns, they can efficiently find and rank the best designs based on specific criteria, like size and motion accuracy. The result is a method for rapidly generating mechanism design concepts, which when integrated in devices could improve accessibility and independence for individuals with movement difficulties. 

A Sit-to-Stand Mobility Assistant device invented by Anurag Purwar, Associate Professor of Mechanical Engineering and his students was brought to the market via a licensing agreement between Stony Brook University and Biodex Medical Systems; see http://www.mobilityassist.net for details. The device was declared as among the top ten inventions ever at Stony Brook University.